Merge tag 'perf-core-for-mingo' of git://git.kernel.org/pub/scm/linux/kernel/git/jolsa/perf into perf/core

Pull perf/core improvements and fixes from Jiri Olsa:

  * Add a test case for hists filtering (Namhyung Kim)

  * Share map_groups among threads of the same group (Arnaldo Carvalho de Melo, Jiri Olsa)

Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>

18 files changed, 948 insertions, 155 deletions

diff --git a/tools/perf/Makefile.perf b/tools/perf/Makefile.perf
index e96923310d57..46e0c32ad23b 100644
--- a/tools/perf/Makefile.perf
+++ b/tools/perf/Makefile.perf
@@ -397,7 +397,9 @@ LIB_OBJS += $(OUTPUT)tests/rdpmc.o
 LIB_OBJS += $(OUTPUT)tests/evsel-roundtrip-name.o
 LIB_OBJS += $(OUTPUT)tests/evsel-tp-sched.o
 LIB_OBJS += $(OUTPUT)tests/pmu.o
+LIB_OBJS += $(OUTPUT)tests/hists_common.o
 LIB_OBJS += $(OUTPUT)tests/hists_link.o
+LIB_OBJS += $(OUTPUT)tests/hists_filter.o
 LIB_OBJS += $(OUTPUT)tests/python-use.o
 LIB_OBJS += $(OUTPUT)tests/bp_signal.o
 LIB_OBJS += $(OUTPUT)tests/bp_signal_overflow.o
@@ -414,6 +416,8 @@ ifeq ($(ARCH),x86)
 LIB_OBJS += $(OUTPUT)tests/dwarf-unwind.o
 endif
 endif
+LIB_OBJS += $(OUTPUT)tests/mmap-thread-lookup.o
+LIB_OBJS += $(OUTPUT)tests/thread-mg-share.o
 
 BUILTIN_OBJS += $(OUTPUT)builtin-annotate.o
 BUILTIN_OBJS += $(OUTPUT)builtin-bench.o
diff --git a/tools/perf/arch/x86/tests/dwarf-unwind.c b/tools/perf/arch/x86/tests/dwarf-unwind.c
index b8c0102c70c8..c916656259f7 100644
--- a/tools/perf/arch/x86/tests/dwarf-unwind.c
+++ b/tools/perf/arch/x86/tests/dwarf-unwind.c
@@ -23,7 +23,7 @@ static int sample_ustack(struct perf_sample *sample,
 
 	sp = (unsigned long) regs[PERF_REG_X86_SP];
 
-	map = map_groups__find(&thread->mg, MAP__FUNCTION, (u64) sp);
+	map = map_groups__find(thread->mg, MAP__FUNCTION, (u64) sp);
 	if (!map) {
 		pr_debug("failed to get stack map\n");
 		free(buf);
diff --git a/tools/perf/perf.h b/tools/perf/perf.h
index 5c11ecad02a9..ebdad3376c67 100644
--- a/tools/perf/perf.h
+++ b/tools/perf/perf.h
@@ -15,6 +15,9 @@
 #ifndef __NR_futex
 # define __NR_futex 240
 #endif
+#ifndef __NR_gettid
+# define __NR_gettid 224
+#endif
 #endif
 
 #if defined(__x86_64__)
@@ -29,6 +32,9 @@
 #ifndef __NR_futex
 # define __NR_futex 202
 #endif
+#ifndef __NR_gettid
+# define __NR_gettid 186
+#endif
 #endif
 
 #ifdef __powerpc__
diff --git a/tools/perf/tests/builtin-test.c b/tools/perf/tests/builtin-test.c
index b11bf8a08430..0d5afaf72944 100644
--- a/tools/perf/tests/builtin-test.c
+++ b/tools/perf/tests/builtin-test.c
@@ -124,6 +124,18 @@ static struct test {
 #endif
 #endif
 	{
+		.desc = "Test filtering hist entries",
+		.func = test__hists_filter,
+	},
+	{
+		.desc = "Test mmap thread lookup",
+		.func = test__mmap_thread_lookup,
+	},
+	{
+		.desc = "Test thread mg sharing",
+		.func = test__thread_mg_share,
+	},
+	{
 		.func = NULL,
 	},
 };
diff --git a/tools/perf/tests/hists_common.c b/tools/perf/tests/hists_common.c
new file mode 100644
index 000000000000..44655b395bb9
--- /dev/null
+++ b/tools/perf/tests/hists_common.c
@@ -0,0 +1,148 @@
+#include "perf.h"
+#include "util/debug.h"
+#include "util/symbol.h"
+#include "util/sort.h"
+#include "util/evsel.h"
+#include "util/evlist.h"
+#include "util/machine.h"
+#include "util/thread.h"
+#include "tests/hists_common.h"
+
+static struct {
+	u32 pid;
+	const char *comm;
+} fake_threads[] = {
+	{ 100, "perf" },
+	{ 200, "perf" },
+	{ 300, "bash" },
+};
+
+static struct {
+	u32 pid;
+	u64 start;
+	const char *filename;
+} fake_mmap_info[] = {
+	{ 100, 0x40000, "perf" },
+	{ 100, 0x50000, "libc" },
+	{ 100, 0xf0000, "[kernel]" },
+	{ 200, 0x40000, "perf" },
+	{ 200, 0x50000, "libc" },
+	{ 200, 0xf0000, "[kernel]" },
+	{ 300, 0x40000, "bash" },
+	{ 300, 0x50000, "libc" },
+	{ 300, 0xf0000, "[kernel]" },
+};
+
+struct fake_sym {
+	u64 start;
+	u64 length;
+	const char *name;
+};
+
+static struct fake_sym perf_syms[] = {
+	{ 700, 100, "main" },
+	{ 800, 100, "run_command" },
+	{ 900, 100, "cmd_record" },
+};
+
+static struct fake_sym bash_syms[] = {
+	{ 700, 100, "main" },
+	{ 800, 100, "xmalloc" },
+	{ 900, 100, "xfree" },
+};
+
+static struct fake_sym libc_syms[] = {
+	{ 700, 100, "malloc" },
+	{ 800, 100, "free" },
+	{ 900, 100, "realloc" },
+};
+
+static struct fake_sym kernel_syms[] = {
+	{ 700, 100, "schedule" },
+	{ 800, 100, "page_fault" },
+	{ 900, 100, "sys_perf_event_open" },
+};
+
+static struct {
+	const char *dso_name;
+	struct fake_sym *syms;
+	size_t nr_syms;
+} fake_symbols[] = {
+	{ "perf", perf_syms, ARRAY_SIZE(perf_syms) },
+	{ "bash", bash_syms, ARRAY_SIZE(bash_syms) },
+	{ "libc", libc_syms, ARRAY_SIZE(libc_syms) },
+	{ "[kernel]", kernel_syms, ARRAY_SIZE(kernel_syms) },
+};
+
+struct machine *setup_fake_machine(struct machines *machines)
+{
+	struct machine *machine = machines__find(machines, HOST_KERNEL_ID);
+	size_t i;
+
+	if (machine == NULL) {
+		pr_debug("Not enough memory for machine setup\n");
+		return NULL;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(fake_threads); i++) {
+		struct thread *thread;
+
+		thread = machine__findnew_thread(machine, fake_threads[i].pid,
+						 fake_threads[i].pid);
+		if (thread == NULL)
+			goto out;
+
+		thread__set_comm(thread, fake_threads[i].comm, 0);
+	}
+
+	for (i = 0; i < ARRAY_SIZE(fake_mmap_info); i++) {
+		union perf_event fake_mmap_event = {
+			.mmap = {
+				.header = { .misc = PERF_RECORD_MISC_USER, },
+				.pid = fake_mmap_info[i].pid,
+				.tid = fake_mmap_info[i].pid,
+				.start = fake_mmap_info[i].start,
+				.len = 0x1000ULL,
+				.pgoff = 0ULL,
+			},
+		};
+
+		strcpy(fake_mmap_event.mmap.filename,
+		       fake_mmap_info[i].filename);
+
+		machine__process_mmap_event(machine, &fake_mmap_event, NULL);
+	}
+
+	for (i = 0; i < ARRAY_SIZE(fake_symbols); i++) {
+		size_t k;
+		struct dso *dso;
+
+		dso = __dsos__findnew(&machine->user_dsos,
+				      fake_symbols[i].dso_name);
+		if (dso == NULL)
+			goto out;
+
+		/* emulate dso__load() */
+		dso__set_loaded(dso, MAP__FUNCTION);
+
+		for (k = 0; k < fake_symbols[i].nr_syms; k++) {
+			struct symbol *sym;
+			struct fake_sym *fsym = &fake_symbols[i].syms[k];
+
+			sym = symbol__new(fsym->start, fsym->length,
+					  STB_GLOBAL, fsym->name);
+			if (sym == NULL)
+				goto out;
+
+			symbols__insert(&dso->symbols[MAP__FUNCTION], sym);
+		}
+	}
+
+	return machine;
+
+out:
+	pr_debug("Not enough memory for machine setup\n");
+	machine__delete_threads(machine);
+	machine__delete(machine);
+	return NULL;
+}
diff --git a/tools/perf/tests/hists_common.h b/tools/perf/tests/hists_common.h
new file mode 100644
index 000000000000..2528b8fc105a
--- /dev/null
+++ b/tools/perf/tests/hists_common.h
@@ -0,0 +1,44 @@
+#ifndef __PERF_TESTS__HISTS_COMMON_H__
+#define __PERF_TESTS__HISTS_COMMON_H__
+
+struct machine;
+struct machines;
+
+/*
+ * The setup_fake_machine() provides a test environment which consists
+ * of 3 processes that have 3 mappings and in turn, have 3 symbols
+ * respectively.  See below table:
+ *
+ * Command:  Pid  Shared Object               Symbol
+ * .............  .............  ...................
+ *    perf:  100           perf  main
+ *    perf:  100           perf  run_command
+ *    perf:  100           perf  comd_record
+ *    perf:  100           libc  malloc
+ *    perf:  100           libc  free
+ *    perf:  100           libc  realloc
+ *    perf:  100       [kernel]  schedule
+ *    perf:  100       [kernel]  page_fault
+ *    perf:  100       [kernel]  sys_perf_event_open
+ *    perf:  200           perf  main
+ *    perf:  200           perf  run_command
+ *    perf:  200           perf  comd_record
+ *    perf:  200           libc  malloc
+ *    perf:  200           libc  free
+ *    perf:  200           libc  realloc
+ *    perf:  200       [kernel]  schedule
+ *    perf:  200       [kernel]  page_fault
+ *    perf:  200       [kernel]  sys_perf_event_open
+ *    bash:  300           bash  main
+ *    bash:  300           bash  xmalloc
+ *    bash:  300           bash  xfree
+ *    bash:  300           libc  malloc
+ *    bash:  300           libc  free
+ *    bash:  300           libc  realloc
+ *    bash:  300       [kernel]  schedule
+ *    bash:  300       [kernel]  page_fault
+ *    bash:  300       [kernel]  sys_perf_event_open
+ */
+struct machine *setup_fake_machine(struct machines *machines);
+
+#endif /* __PERF_TESTS__HISTS_COMMON_H__ */
diff --git a/tools/perf/tests/hists_filter.c b/tools/perf/tests/hists_filter.c
new file mode 100644
index 000000000000..23dc2f4d12c3
--- /dev/null
+++ b/tools/perf/tests/hists_filter.c
@@ -0,0 +1,315 @@
+#include "perf.h"
+#include "util/debug.h"
+#include "util/symbol.h"
+#include "util/sort.h"
+#include "util/evsel.h"
+#include "util/evlist.h"
+#include "util/machine.h"
+#include "util/thread.h"
+#include "util/parse-events.h"
+#include "tests/tests.h"
+#include "tests/hists_common.h"
+
+struct sample {
+	u32 pid;
+	u64 ip;
+	struct thread *thread;
+	struct map *map;
+	struct symbol *sym;
+};
+
+/* For the numbers, see hists_common.c */
+static struct sample fake_samples[] = {
+	/* perf [kernel] schedule() */
+	{ .pid = 100, .ip = 0xf0000 + 700, },
+	/* perf [perf]   main() */
+	{ .pid = 100, .ip = 0x40000 + 700, },
+	/* perf [libc]   malloc() */
+	{ .pid = 100, .ip = 0x50000 + 700, },
+	/* perf [perf]   main() */
+	{ .pid = 200, .ip = 0x40000 + 700, }, /* will be merged */
+	/* perf [perf]   cmd_record() */
+	{ .pid = 200, .ip = 0x40000 + 900, },
+	/* perf [kernel] page_fault() */
+	{ .pid = 200, .ip = 0xf0000 + 800, },
+	/* bash [bash]   main() */
+	{ .pid = 300, .ip = 0x40000 + 700, },
+	/* bash [bash]   xmalloc() */
+	{ .pid = 300, .ip = 0x40000 + 800, },
+	/* bash [libc]   malloc() */
+	{ .pid = 300, .ip = 0x50000 + 700, },
+	/* bash [kernel] page_fault() */
+	{ .pid = 300, .ip = 0xf0000 + 800, },
+};
+
+static int add_hist_entries(struct perf_evlist *evlist, struct machine *machine)
+{
+	struct perf_evsel *evsel;
+	struct addr_location al;
+	struct hist_entry *he;
+	struct perf_sample sample = { .cpu = 0, };
+	size_t i;
+
+	/*
+	 * each evsel will have 10 samples but the 4th sample
+	 * (perf [perf] main) will be collapsed to an existing entry
+	 * so total 9 entries will be in the tree.
+	 */
+	evlist__for_each(evlist, evsel) {
+		for (i = 0; i < ARRAY_SIZE(fake_samples); i++) {
+			const union perf_event event = {
+				.header = {
+					.misc = PERF_RECORD_MISC_USER,
+				},
+			};
+
+			/* make sure it has no filter at first */
+			evsel->hists.thread_filter = NULL;
+			evsel->hists.dso_filter = NULL;
+			evsel->hists.symbol_filter_str = NULL;
+
+			sample.pid = fake_samples[i].pid;
+			sample.ip = fake_samples[i].ip;
+
+			if (perf_event__preprocess_sample(&event, machine, &al,
+							  &sample) < 0)
+				goto out;
+
+			he = __hists__add_entry(&evsel->hists, &al, NULL,
+						NULL, NULL, 100, 1, 0);
+			if (he == NULL)
+				goto out;
+
+			fake_samples[i].thread = al.thread;
+			fake_samples[i].map = al.map;
+			fake_samples[i].sym = al.sym;
+
+			hists__inc_nr_events(he->hists, PERF_RECORD_SAMPLE);
+			if (!he->filtered)
+				he->hists->stats.nr_non_filtered_samples++;
+		}
+	}
+
+	return 0;
+
+out:
+	pr_debug("Not enough memory for adding a hist entry\n");
+	return TEST_FAIL;
+}
+
+static void print_hists(struct hists *hists)
+{
+	int i = 0;
+	struct rb_root *root;
+	struct rb_node *node;
+
+	root = &hists->entries;
+
+	pr_info("----- %s --------\n", __func__);
+	node = rb_first(root);
+	while (node) {
+		struct hist_entry *he;
+
+		he = rb_entry(node, struct hist_entry, rb_node);
+
+		if (!he->filtered) {
+			pr_info("%2d: entry: %-8s [%-8s] %20s: period = %"PRIu64"\n",
+				i, thread__comm_str(he->thread),
+				he->ms.map->dso->short_name,
+				he->ms.sym->name, he->stat.period);
+		}
+
+		i++;
+		node = rb_next(node);
+	}
+}
+
+int test__hists_filter(void)
+{
+	int err = TEST_FAIL;
+	struct machines machines;
+	struct machine *machine;
+	struct perf_evsel *evsel;
+	struct perf_evlist *evlist = perf_evlist__new();
+
+	TEST_ASSERT_VAL("No memory", evlist);
+
+	err = parse_events(evlist, "cpu-clock");
+	if (err)
+		goto out;
+	err = parse_events(evlist, "task-clock");
+	if (err)
+		goto out;
+
+	/* default sort order (comm,dso,sym) will be used */
+	if (setup_sorting() < 0)
+		goto out;
+
+	machines__init(&machines);
+
+	/* setup threads/dso/map/symbols also */
+	machine = setup_fake_machine(&machines);
+	if (!machine)
+		goto out;
+
+	if (verbose > 1)
+		machine__fprintf(machine, stderr);
+
+	/* process sample events */
+	err = add_hist_entries(evlist, machine);
+	if (err < 0)
+		goto out;
+
+	evlist__for_each(evlist, evsel) {
+		struct hists *hists = &evsel->hists;
+
+		hists__collapse_resort(hists, NULL);
+		hists__output_resort(hists);
+
+		if (verbose > 2) {
+			pr_info("Normal histogram\n");
+			print_hists(hists);
+		}
+
+		TEST_ASSERT_VAL("Invalid nr samples",
+				hists->stats.nr_events[PERF_RECORD_SAMPLE] == 10);
+		TEST_ASSERT_VAL("Invalid nr hist entries",
+				hists->nr_entries == 9);
+		TEST_ASSERT_VAL("Invalid total period",
+				hists->stats.total_period == 1000);
+		TEST_ASSERT_VAL("Unmatched nr samples",
+				hists->stats.nr_events[PERF_RECORD_SAMPLE] ==
+				hists->stats.nr_non_filtered_samples);
+		TEST_ASSERT_VAL("Unmatched nr hist entries",
+				hists->nr_entries == hists->nr_non_filtered_entries);
+		TEST_ASSERT_VAL("Unmatched total period",
+				hists->stats.total_period ==
+				hists->stats.total_non_filtered_period);
+
+		/* now applying thread filter for 'bash' */
+		evsel->hists.thread_filter = fake_samples[9].thread;
+		hists__filter_by_thread(hists);
+
+		if (verbose > 2) {
+			pr_info("Histogram for thread filter\n");
+			print_hists(hists);
+		}
+
+		/* normal stats should be invariant */
+		TEST_ASSERT_VAL("Invalid nr samples",
+				hists->stats.nr_events[PERF_RECORD_SAMPLE] == 10);
+		TEST_ASSERT_VAL("Invalid nr hist entries",
+				hists->nr_entries == 9);
+		TEST_ASSERT_VAL("Invalid total period",
+				hists->stats.total_period == 1000);
+
+		/* but filter stats are changed */
+		TEST_ASSERT_VAL("Unmatched nr samples for thread filter",
+				hists->stats.nr_non_filtered_samples == 4);
+		TEST_ASSERT_VAL("Unmatched nr hist entries for thread filter",
+				hists->nr_non_filtered_entries == 4);
+		TEST_ASSERT_VAL("Unmatched total period for thread filter",
+				hists->stats.total_non_filtered_period == 400);
+
+		/* remove thread filter first */
+		evsel->hists.thread_filter = NULL;
+		hists__filter_by_thread(hists);
+
+		/* now applying dso filter for 'kernel' */
+		evsel->hists.dso_filter = fake_samples[0].map->dso;
+		hists__filter_by_dso(hists);
+
+		if (verbose > 2) {
+			pr_info("Histogram for dso filter\n");
+			print_hists(hists);
+		}
+
+		/* normal stats should be invariant */
+		TEST_ASSERT_VAL("Invalid nr samples",
+				hists->stats.nr_events[PERF_RECORD_SAMPLE] == 10);
+		TEST_ASSERT_VAL("Invalid nr hist entries",
+				hists->nr_entries == 9);
+		TEST_ASSERT_VAL("Invalid total period",
+				hists->stats.total_period == 1000);
+
+		/* but filter stats are changed */
+		TEST_ASSERT_VAL("Unmatched nr samples for dso filter",
+				hists->stats.nr_non_filtered_samples == 3);
+		TEST_ASSERT_VAL("Unmatched nr hist entries for dso filter",
+				hists->nr_non_filtered_entries == 3);
+		TEST_ASSERT_VAL("Unmatched total period for dso filter",
+				hists->stats.total_non_filtered_period == 300);
+
+		/* remove dso filter first */
+		evsel->hists.dso_filter = NULL;
+		hists__filter_by_dso(hists);
+
+		/*
+		 * now applying symbol filter for 'main'.  Also note that
+		 * there's 3 samples that have 'main' symbol but the 4th
+		 * entry of fake_samples was collapsed already so it won't
+		 * be counted as a separate entry but the sample count and
+		 * total period will be remained.
+		 */
+		evsel->hists.symbol_filter_str = "main";
+		hists__filter_by_symbol(hists);
+
+		if (verbose > 2) {
+			pr_info("Histogram for symbol filter\n");
+			print_hists(hists);
+		}
+
+		/* normal stats should be invariant */
+		TEST_ASSERT_VAL("Invalid nr samples",
+				hists->stats.nr_events[PERF_RECORD_SAMPLE] == 10);
+		TEST_ASSERT_VAL("Invalid nr hist entries",
+				hists->nr_entries == 9);
+		TEST_ASSERT_VAL("Invalid total period",
+				hists->stats.total_period == 1000);
+
+		/* but filter stats are changed */
+		TEST_ASSERT_VAL("Unmatched nr samples for symbol filter",
+				hists->stats.nr_non_filtered_samples == 3);
+		TEST_ASSERT_VAL("Unmatched nr hist entries for symbol filter",
+				hists->nr_non_filtered_entries == 2);
+		TEST_ASSERT_VAL("Unmatched total period for symbol filter",
+				hists->stats.total_non_filtered_period == 300);
+
+		/* now applying all filters at once. */
+		evsel->hists.thread_filter = fake_samples[1].thread;
+		evsel->hists.dso_filter = fake_samples[1].map->dso;
+		hists__filter_by_thread(hists);
+		hists__filter_by_dso(hists);
+
+		if (verbose > 2) {
+			pr_info("Histogram for all filters\n");
+			print_hists(hists);
+		}
+
+		/* normal stats should be invariant */
+		TEST_ASSERT_VAL("Invalid nr samples",
+				hists->stats.nr_events[PERF_RECORD_SAMPLE] == 10);
+		TEST_ASSERT_VAL("Invalid nr hist entries",
+				hists->nr_entries == 9);
+		TEST_ASSERT_VAL("Invalid total period",
+				hists->stats.total_period == 1000);
+
+		/* but filter stats are changed */
+		TEST_ASSERT_VAL("Unmatched nr samples for all filter",
+				hists->stats.nr_non_filtered_samples == 2);
+		TEST_ASSERT_VAL("Unmatched nr hist entries for all filter",
+				hists->nr_non_filtered_entries == 1);
+		TEST_ASSERT_VAL("Unmatched total period for all filter",
+				hists->stats.total_non_filtered_period == 200);
+	}
+
+
+	err = TEST_OK;
+
+out:
+	/* tear down everything */
+	perf_evlist__delete(evlist);
+	machines__exit(&machines);
+
+	return err;
+}
diff --git a/tools/perf/tests/hists_link.c b/tools/perf/tests/hists_link.c
index 7ccbc7b6ae77..e42d6790811a 100644
--- a/tools/perf/tests/hists_link.c
+++ b/tools/perf/tests/hists_link.c
@@ -8,145 +8,7 @@
 #include "machine.h"
 #include "thread.h"
 #include "parse-events.h"
-
-static struct {
-	u32 pid;
-	const char *comm;
-} fake_threads[] = {
-	{ 100, "perf" },
-	{ 200, "perf" },
-	{ 300, "bash" },
-};
-
-static struct {
-	u32 pid;
-	u64 start;
-	const char *filename;
-} fake_mmap_info[] = {
-	{ 100, 0x40000, "perf" },
-	{ 100, 0x50000, "libc" },
-	{ 100, 0xf0000, "[kernel]" },
-	{ 200, 0x40000, "perf" },
-	{ 200, 0x50000, "libc" },
-	{ 200, 0xf0000, "[kernel]" },
-	{ 300, 0x40000, "bash" },
-	{ 300, 0x50000, "libc" },
-	{ 300, 0xf0000, "[kernel]" },
-};
-
-struct fake_sym {
-	u64 start;
-	u64 length;
-	const char *name;
-};
-
-static struct fake_sym perf_syms[] = {
-	{ 700, 100, "main" },
-	{ 800, 100, "run_command" },
-	{ 900, 100, "cmd_record" },
-};
-
-static struct fake_sym bash_syms[] = {
-	{ 700, 100, "main" },
-	{ 800, 100, "xmalloc" },
-	{ 900, 100, "xfree" },
-};
-
-static struct fake_sym libc_syms[] = {
-	{ 700, 100, "malloc" },
-	{ 800, 100, "free" },
-	{ 900, 100, "realloc" },
-};
-
-static struct fake_sym kernel_syms[] = {
-	{ 700, 100, "schedule" },
-	{ 800, 100, "page_fault" },
-	{ 900, 100, "sys_perf_event_open" },
-};
-
-static struct {
-	const char *dso_name;
-	struct fake_sym *syms;
-	size_t nr_syms;
-} fake_symbols[] = {
-	{ "perf", perf_syms, ARRAY_SIZE(perf_syms) },
-	{ "bash", bash_syms, ARRAY_SIZE(bash_syms) },
-	{ "libc", libc_syms, ARRAY_SIZE(libc_syms) },
-	{ "[kernel]", kernel_syms, ARRAY_SIZE(kernel_syms) },
-};
-
-static struct machine *setup_fake_machine(struct machines *machines)
-{
-	struct machine *machine = machines__find(machines, HOST_KERNEL_ID);
-	size_t i;
-
-	if (machine == NULL) {
-		pr_debug("Not enough memory for machine setup\n");
-		return NULL;
-	}
-
-	for (i = 0; i < ARRAY_SIZE(fake_threads); i++) {
-		struct thread *thread;
-
-		thread = machine__findnew_thread(machine, fake_threads[i].pid,
-						 fake_threads[i].pid);
-		if (thread == NULL)
-			goto out;
-
-		thread__set_comm(thread, fake_threads[i].comm, 0);
-	}
-
-	for (i = 0; i < ARRAY_SIZE(fake_mmap_info); i++) {
-		union perf_event fake_mmap_event = {
-			.mmap = {
-				.header = { .misc = PERF_RECORD_MISC_USER, },
-				.pid = fake_mmap_info[i].pid,
-				.tid = fake_mmap_info[i].pid,
-				.start = fake_mmap_info[i].start,
-				.len = 0x1000ULL,
-				.pgoff = 0ULL,
-			},
-		};
-
-		strcpy(fake_mmap_event.mmap.filename,
-		       fake_mmap_info[i].filename);
-
-		machine__process_mmap_event(machine, &fake_mmap_event, NULL);
-	}
-
-	for (i = 0; i < ARRAY_SIZE(fake_symbols); i++) {
-		size_t k;
-		struct dso *dso;
-
-		dso = __dsos__findnew(&machine->user_dsos,
-				      fake_symbols[i].dso_name);
-		if (dso == NULL)
-			goto out;
-
-		/* emulate dso__load() */
-		dso__set_loaded(dso, MAP__FUNCTION);
-
-		for (k = 0; k < fake_symbols[i].nr_syms; k++) {
-			struct symbol *sym;
-			struct fake_sym *fsym = &fake_symbols[i].syms[k];
-
-			sym = symbol__new(fsym->start, fsym->length,
-					  STB_GLOBAL, fsym->name);
-			if (sym == NULL)
-				goto out;
-
-			symbols__insert(&dso->symbols[MAP__FUNCTION], sym);
-		}
-	}
-
-	return machine;
-
-out:
-	pr_debug("Not enough memory for machine setup\n");
-	machine__delete_threads(machine);
-	machine__delete(machine);
-	return NULL;
-}
+#include "hists_common.h"
 
 struct sample {
 	u32 pid;
@@ -156,6 +18,7 @@ struct sample {
 	struct symbol *sym;
 };
 
+/* For the numbers, see hists_common.c */
 static struct sample fake_common_samples[] = {
 	/* perf [kernel] schedule() */
 	{ .pid = 100, .ip = 0xf0000 + 700, },
diff --git a/tools/perf/tests/mmap-thread-lookup.c b/tools/perf/tests/mmap-thread-lookup.c
new file mode 100644
index 000000000000..4a456fef66ca
--- /dev/null
+++ b/tools/perf/tests/mmap-thread-lookup.c
@@ -0,0 +1,233 @@
+#include <unistd.h>
+#include <sys/syscall.h>
+#include <sys/types.h>
+#include <sys/mman.h>
+#include <pthread.h>
+#include <stdlib.h>
+#include <stdio.h>
+#include "debug.h"
+#include "tests.h"
+#include "machine.h"
+#include "thread_map.h"
+#include "symbol.h"
+#include "thread.h"
+
+#define THREADS 4
+
+static int go_away;
+
+struct thread_data {
+	pthread_t	pt;
+	pid_t		tid;
+	void		*map;
+	int		ready[2];
+};
+
+static struct thread_data threads[THREADS];
+
+static int thread_init(struct thread_data *td)
+{
+	void *map;
+
+	map = mmap(NULL, page_size,
+		   PROT_READ|PROT_WRITE|PROT_EXEC,
+		   MAP_SHARED|MAP_ANONYMOUS, -1, 0);
+
+	if (map == MAP_FAILED) {
+		perror("mmap failed");
+		return -1;
+	}
+
+	td->map = map;
+	td->tid = syscall(SYS_gettid);
+
+	pr_debug("tid = %d, map = %p\n", td->tid, map);
+	return 0;
+}
+
+static void *thread_fn(void *arg)
+{
+	struct thread_data *td = arg;
+	ssize_t ret;
+	int go;
+
+	if (thread_init(td))
+		return NULL;
+
+	/* Signal thread_create thread is initialized. */
+	ret = write(td->ready[1], &go, sizeof(int));
+	if (ret != sizeof(int)) {
+		pr_err("failed to notify\n");
+		return NULL;
+	}
+
+	while (!go_away) {
+		/* Waiting for main thread to kill us. */
+		usleep(100);
+	}
+
+	munmap(td->map, page_size);
+	return NULL;
+}
+
+static int thread_create(int i)
+{
+	struct thread_data *td = &threads[i];
+	int err, go;
+
+	if (pipe(td->ready))
+		return -1;
+
+	err = pthread_create(&td->pt, NULL, thread_fn, td);
+	if (!err) {
+		/* Wait for thread initialization. */
+		ssize_t ret = read(td->ready[0], &go, sizeof(int));
+		err = ret != sizeof(int);
+	}
+
+	close(td->ready[0]);
+	close(td->ready[1]);
+	return err;
+}
+
+static int threads_create(void)
+{
+	struct thread_data *td0 = &threads[0];
+	int i, err = 0;
+
+	go_away = 0;
+
+	/* 0 is main thread */
+	if (thread_init(td0))
+		return -1;
+
+	for (i = 1; !err && i < THREADS; i++)
+		err = thread_create(i);
+
+	return err;
+}
+
+static int threads_destroy(void)
+{
+	struct thread_data *td0 = &threads[0];
+	int i, err = 0;
+
+	/* cleanup the main thread */
+	munmap(td0->map, page_size);
+
+	go_away = 1;
+
+	for (i = 1; !err && i < THREADS; i++)
+		err = pthread_join(threads[i].pt, NULL);
+
+	return err;
+}
+
+typedef int (*synth_cb)(struct machine *machine);
+
+static int synth_all(struct machine *machine)
+{
+	return perf_event__synthesize_threads(NULL,
+					      perf_event__process,
+					      machine, 0);
+}
+
+static int synth_process(struct machine *machine)
+{
+	struct thread_map *map;
+	int err;
+
+	map = thread_map__new_by_pid(getpid());
+
+	err = perf_event__synthesize_thread_map(NULL, map,
+						perf_event__process,
+						machine, 0);
+
+	thread_map__delete(map);
+	return err;
+}
+
+static int mmap_events(synth_cb synth)
+{
+	struct machines machines;
+	struct machine *machine;
+	int err, i;
+
+	/*
+	 * The threads_create will not return before all threads
+	 * are spawned and all created memory map.
+	 *
+	 * They will loop until threads_destroy is called, so we
+	 * can safely run synthesizing function.
+	 */
+	TEST_ASSERT_VAL("failed to create threads", !threads_create());
+
+	machines__init(&machines);
+	machine = &machines.host;
+
+	dump_trace = verbose > 1 ? 1 : 0;
+
+	err = synth(machine);
+
+	dump_trace = 0;
+
+	TEST_ASSERT_VAL("failed to destroy threads", !threads_destroy());
+	TEST_ASSERT_VAL("failed to synthesize maps", !err);
+
+	/*
+	 * All data is synthesized, try to find map for each
+	 * thread object.
+	 */
+	for (i = 0; i < THREADS; i++) {
+		struct thread_data *td = &threads[i];
+		struct addr_location al;
+		struct thread *thread;
+
+		thread = machine__findnew_thread(machine, getpid(), td->tid);
+
+		pr_debug("looking for map %p\n", td->map);
+
+		thread__find_addr_map(thread, machine,
+				      PERF_RECORD_MISC_USER, MAP__FUNCTION,
+				      (unsigned long) (td->map + 1), &al);
+
+		if (!al.map) {
+			pr_debug("failed, couldn't find map\n");
+			err = -1;
+			break;
+		}
+
+		pr_debug("map %p, addr %" PRIx64 "\n", al.map, al.map->start);
+	}
+
+	machine__delete_threads(machine);
+	machines__exit(&machines);
+	return err;
+}
+
+/*
+ * This test creates 'THREADS' number of threads (including
+ * main thread) and each thread creates memory map.
+ *
+ * When threads are created, we synthesize them with both
+ * (separate tests):
+ *   perf_event__synthesize_thread_map (process based)
+ *   perf_event__synthesize_threads    (global)
+ *
+ * We test we can find all memory maps via:
+ *   thread__find_addr_map
+ *
+ * by using all thread objects.
+ */
+int test__mmap_thread_lookup(void)
+{
+	/* perf_event__synthesize_threads synthesize */
+	TEST_ASSERT_VAL("failed with sythesizing all",
+			!mmap_events(synth_all));
+
+	/* perf_event__synthesize_thread_map synthesize */
+	TEST_ASSERT_VAL("failed with sythesizing process",
+			!mmap_events(synth_process));
+
+	return 0;
+}
diff --git a/tools/perf/tests/tests.h b/tools/perf/tests/tests.h
index a24795ca002d..a9d7cb019f9e 100644
--- a/tools/perf/tests/tests.h
+++ b/tools/perf/tests/tests.h
@@ -41,6 +41,9 @@ int test__sample_parsing(void);
 int test__keep_tracking(void);
 int test__parse_no_sample_id_all(void);
 int test__dwarf_unwind(void);
+int test__hists_filter(void);
+int test__mmap_thread_lookup(void);
+int test__thread_mg_share(void);
 
 #if defined(__x86_64__) || defined(__i386__)
 #ifdef HAVE_DWARF_UNWIND_SUPPORT
diff --git a/tools/perf/tests/thread-mg-share.c b/tools/perf/tests/thread-mg-share.c
new file mode 100644
index 000000000000..2b2e0dbe114f
--- /dev/null
+++ b/tools/perf/tests/thread-mg-share.c
@@ -0,0 +1,90 @@
+#include "tests.h"
+#include "machine.h"
+#include "thread.h"
+#include "map.h"
+
+int test__thread_mg_share(void)
+{
+	struct machines machines;
+	struct machine *machine;
+
+	/* thread group */
+	struct thread *leader;
+	struct thread *t1, *t2, *t3;
+	struct map_groups *mg;
+
+	/* other process */
+	struct thread *other, *other_leader;
+	struct map_groups *other_mg;
+
+	/*
+	 * This test create 2 processes abstractions (struct thread)
+	 * with several threads and checks they properly share and
+	 * maintain map groups info (struct map_groups).
+	 *
+	 * thread group (pid: 0, tids: 0, 1, 2, 3)
+	 * other  group (pid: 4, tids: 4, 5)
+	*/
+
+	machines__init(&machines);
+	machine = &machines.host;
+
+	/* create process with 4 threads */
+	leader = machine__findnew_thread(machine, 0, 0);
+	t1     = machine__findnew_thread(machine, 0, 1);
+	t2     = machine__findnew_thread(machine, 0, 2);
+	t3     = machine__findnew_thread(machine, 0, 3);
+
+	/* and create 1 separated process, without thread leader */
+	other  = machine__findnew_thread(machine, 4, 5);
+
+	TEST_ASSERT_VAL("failed to create threads",
+			leader && t1 && t2 && t3 && other);
+
+	mg = leader->mg;
+	TEST_ASSERT_VAL("wrong refcnt", mg->refcnt == 4);
+
+	/* test the map groups pointer is shared */
+	TEST_ASSERT_VAL("map groups don't match", mg == t1->mg);
+	TEST_ASSERT_VAL("map groups don't match", mg == t2->mg);
+	TEST_ASSERT_VAL("map groups don't match", mg == t3->mg);
+
+	/*
+	 * Verify the other leader was created by previous call.
+	 * It should have shared map groups with no change in
+	 * refcnt.
+	 */
+	other_leader = machine__find_thread(machine, 4, 4);
+	TEST_ASSERT_VAL("failed to find other leader", other_leader);
+
+	other_mg = other->mg;
+	TEST_ASSERT_VAL("wrong refcnt", other_mg->refcnt == 2);
+
+	TEST_ASSERT_VAL("map groups don't match", other_mg == other_leader->mg);
+
+	/* release thread group */
+	thread__delete(leader);
+	TEST_ASSERT_VAL("wrong refcnt", mg->refcnt == 3);
+
+	thread__delete(t1);
+	TEST_ASSERT_VAL("wrong refcnt", mg->refcnt == 2);
+
+	thread__delete(t2);
+	TEST_ASSERT_VAL("wrong refcnt", mg->refcnt == 1);
+
+	thread__delete(t3);
+
+	/* release other group  */
+	thread__delete(other_leader);
+	TEST_ASSERT_VAL("wrong refcnt", other_mg->refcnt == 1);
+
+	thread__delete(other);
+
+	/*
+	 * Cannot call machine__delete_threads(machine) now,
+	 * because we've already released all the threads.
+	 */
+
+	machines__exit(&machines);
+	return 0;
+}
diff --git a/tools/perf/ui/stdio/hist.c b/tools/perf/ui/stdio/hist.c
index d59893edf031..9eccf7f4f367 100644
--- a/tools/perf/ui/stdio/hist.c
+++ b/tools/perf/ui/stdio/hist.c
@@ -495,7 +495,7 @@ print_entries:
 			break;
 
 		if (h->ms.map == NULL && verbose > 1) {
-			__map_groups__fprintf_maps(&h->thread->mg,
+			__map_groups__fprintf_maps(h->thread->mg,
 						   MAP__FUNCTION, verbose, fp);
 			fprintf(fp, "%.10s end\n", graph_dotted_line);
 		}
diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c
index 9d12aa6dd485..dbcaea1a8180 100644
--- a/tools/perf/util/event.c
+++ b/tools/perf/util/event.c
@@ -699,7 +699,7 @@ void thread__find_addr_map(struct thread *thread,
 			   enum map_type type, u64 addr,
 			   struct addr_location *al)
 {
-	struct map_groups *mg = &thread->mg;
+	struct map_groups *mg = thread->mg;
 	bool load_map = false;
 
 	al->machine = machine;
diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c
index a53cd0b8c151..98ec56dc890b 100644
--- a/tools/perf/util/machine.c
+++ b/tools/perf/util/machine.c
@@ -316,6 +316,17 @@ static struct thread *__machine__findnew_thread(struct machine *machine,
 		rb_link_node(&th->rb_node, parent, p);
 		rb_insert_color(&th->rb_node, &machine->threads);
 		machine->last_match = th;
+
+		/*
+		 * We have to initialize map_groups separately
+		 * after rb tree is updated.
+		 *
+		 * The reason is that we call machine__findnew_thread
+		 * within thread__init_map_groups to find the thread
+		 * leader and that would screwed the rb tree.
+		 */
+		if (thread__init_map_groups(th, machine))
+			return NULL;
 	}
 
 	return th;
diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c
index 39cd2d0faff6..ba5f5c0c838b 100644
--- a/tools/perf/util/map.c
+++ b/tools/perf/util/map.c
@@ -323,6 +323,7 @@ void map_groups__init(struct map_groups *mg)
 		INIT_LIST_HEAD(&mg->removed_maps[i]);
 	}
 	mg->machine = NULL;
+	mg->refcnt = 1;
 }
 
 static void maps__delete(struct rb_root *maps)
@@ -358,6 +359,28 @@ void map_groups__exit(struct map_groups *mg)
 	}
 }
 
+struct map_groups *map_groups__new(void)
+{
+	struct map_groups *mg = malloc(sizeof(*mg));
+
+	if (mg != NULL)
+		map_groups__init(mg);
+
+	return mg;
+}
+
+void map_groups__delete(struct map_groups *mg)
+{
+	map_groups__exit(mg);
+	free(mg);
+}
+
+void map_groups__put(struct map_groups *mg)
+{
+	if (--mg->refcnt == 0)
+		map_groups__delete(mg);
+}
+
 void map_groups__flush(struct map_groups *mg)
 {
 	int type;
diff --git a/tools/perf/util/map.h b/tools/perf/util/map.h
index f00f058afb3b..d6445b27d672 100644
--- a/tools/perf/util/map.h
+++ b/tools/perf/util/map.h
@@ -59,8 +59,20 @@ struct map_groups {
 	struct rb_root	 maps[MAP__NR_TYPES];
 	struct list_head removed_maps[MAP__NR_TYPES];
 	struct machine	 *machine;
+	int		 refcnt;
 };
 
+struct map_groups *map_groups__new(void);
+void map_groups__delete(struct map_groups *mg);
+
+static inline struct map_groups *map_groups__get(struct map_groups *mg)
+{
+	++mg->refcnt;
+	return mg;
+}
+
+void map_groups__put(struct map_groups *mg);
+
 static inline struct kmap *map__kmap(struct map *map)
 {
 	return (struct kmap *)(map + 1);
diff --git a/tools/perf/util/thread.c b/tools/perf/util/thread.c
index 3ce0498bdae6..2fde0d5e40b5 100644
--- a/tools/perf/util/thread.c
+++ b/tools/perf/util/thread.c
@@ -8,6 +8,22 @@
 #include "debug.h"
 #include "comm.h"
 
+int thread__init_map_groups(struct thread *thread, struct machine *machine)
+{
+	struct thread *leader;
+	pid_t pid = thread->pid_;
+
+	if (pid == thread->tid) {
+		thread->mg = map_groups__new();
+	} else {
+		leader = machine__findnew_thread(machine, pid, pid);
+		if (leader)
+			thread->mg = map_groups__get(leader->mg);
+	}
+
+	return thread->mg ? 0 : -1;
+}
+
 struct thread *thread__new(pid_t pid, pid_t tid)
 {
 	char *comm_str;
@@ -15,7 +31,6 @@ struct thread *thread__new(pid_t pid, pid_t tid)
 	struct thread *thread = zalloc(sizeof(*thread));
 
 	if (thread != NULL) {
-		map_groups__init(&thread->mg);
 		thread->pid_ = pid;
 		thread->tid = tid;
 		thread->ppid = -1;
@@ -45,7 +60,8 @@ void thread__delete(struct thread *thread)
 {
 	struct comm *comm, *tmp;
 
-	map_groups__exit(&thread->mg);
+	map_groups__put(thread->mg);
+	thread->mg = NULL;
 	list_for_each_entry_safe(comm, tmp, &thread->comm_list, list) {
 		list_del(&comm->list);
 		comm__free(comm);
@@ -111,18 +127,35 @@ int thread__comm_len(struct thread *thread)
 size_t thread__fprintf(struct thread *thread, FILE *fp)
 {
 	return fprintf(fp, "Thread %d %s\n", thread->tid, thread__comm_str(thread)) +
-	       map_groups__fprintf(&thread->mg, verbose, fp);
+	       map_groups__fprintf(thread->mg, verbose, fp);
 }
 
 void thread__insert_map(struct thread *thread, struct map *map)
 {
-	map_groups__fixup_overlappings(&thread->mg, map, verbose, stderr);
-	map_groups__insert(&thread->mg, map);
+	map_groups__fixup_overlappings(thread->mg, map, verbose, stderr);
+	map_groups__insert(thread->mg, map);
+}
+
+static int thread__clone_map_groups(struct thread *thread,
+				    struct thread *parent)
+{
+	int i;
+
+	/* This is new thread, we share map groups for process. */
+	if (thread->pid_ == parent->pid_)
+		return 0;
+
+	/* But this one is new process, copy maps. */
+	for (i = 0; i < MAP__NR_TYPES; ++i)
+		if (map_groups__clone(thread->mg, parent->mg, i) < 0)
+			return -ENOMEM;
+
+	return 0;
 }
 
 int thread__fork(struct thread *thread, struct thread *parent, u64 timestamp)
 {
-	int i, err;
+	int err;
 
 	if (parent->comm_set) {
 		const char *comm = thread__comm_str(parent);
@@ -134,13 +167,8 @@ int thread__fork(struct thread *thread, struct thread *parent, u64 timestamp)
 		thread->comm_set = true;
 	}
 
-	for (i = 0; i < MAP__NR_TYPES; ++i)
-		if (map_groups__clone(&thread->mg, &parent->mg, i) < 0)
-			return -ENOMEM;
-
 	thread->ppid = parent->tid;
-
-	return 0;
+	return thread__clone_map_groups(thread, parent);
 }
 
 void thread__find_cpumode_addr_location(struct thread *thread,
diff --git a/tools/perf/util/thread.h b/tools/perf/util/thread.h
index 9b29f085aede..3c0c2724f82c 100644
--- a/tools/perf/util/thread.h
+++ b/tools/perf/util/thread.h
@@ -13,7 +13,7 @@ struct thread {
 		struct rb_node	 rb_node;
 		struct list_head node;
 	};
-	struct map_groups	mg;
+	struct map_groups	*mg;
 	pid_t			pid_; /* Not all tools update this */
 	pid_t			tid;
 	pid_t			ppid;
@@ -30,6 +30,7 @@ struct machine;
 struct comm;
 
 struct thread *thread__new(pid_t pid, pid_t tid);
+int thread__init_map_groups(struct thread *thread, struct machine *machine);
 void thread__delete(struct thread *thread);
 static inline void thread__exited(struct thread *thread)
 {